# MARKA AI
Autonomous AI Marketing Automation Framework focused on an Arabic-first experience for Egyptian SMEs.
## 🚀 Quick Start (Local Development)
This project uses `uv` for Python package management and Docker Compose for running the infrastructure.
### Prerequisites
- Docker & Docker Compose
- uv (Python package manager)
### Setup
1. **Environment Variables**
Copy the example environment file and fill in your API keys (Gemini, Telegram):
```bash
cp .env.example .env
```
2. **Start the Infrastructure**
This will spin up MongoDB, Qdrant, the FastAPI backend, and dummy containers for the bot and web UI.
```bash
docker compose up --build
```
3. **Access the API**
Once the health checks pass, FastAPI will be available at:
- Docs:
localhost
- Health:
localhost
## 🏗 Architecture
- **AI Layer:** FastAPI, LiteLLM, Pydantic.
- **Databases:** MongoDB 8.0 (State), Qdrant v1.17.0 (Vector Store).
- **LLMs:** Google Gemini (Smart Orchestrator Strategy using `gemini-3-flash` and `gemini-3.1-pro`).
See `MARKA_AI_Technical_Document.md` for full architectural details.